SOC200H1—Measurement October 1 2012
Outline of todays Class:
This lecture will focus on the ideas of how x causes y, how x affects y, and under what
conditions. It will also focus on the measurement of concepts, units, and levels.
How does X cause Y?
To begin, what is a mechanism? It is an account of how, or by the process, one thing
causes another. For example, people with more education are seen as more skilled by potential
employers and get higher status jobs that pay more money. Another example is that people with
more education have more and better employment options, and are therefore able to negotiate
An example of a causal relationship would be how working fewer hours for pay causes
better grades. The mechanism of this idea would be that people who work have fewer hours to
spend studying due to work. This is compared to those who don’t work and can spend more time
Another causal relationship could be the idea that living in a city it better for peoples
health. The mechanism of this could be that people who live in cities are more likely to have
better immune systems than those living in rural areas because they have more networks.
A third example of a causal relationship is that having kids causes less sleep. The
mechanism here would be that people with children may be more worried about their children,
thus retaining less sleep per night. Another mechanism to this example would be that people
who have class may work longer hours to support their children, and therefore, they also get less
Therefore, we can conclude that a proposed mechanism is a theory of how X leads to Y.
This leads us to the theory of credentialism. This is the idea that people with more education
have more credentials than people with less education. Credentials identify that the candidate
has competence or other desirable attribute an employer wants. People with credentials are able
to get jobs that pay more money than people who don’t have them. Therefore, credentials signal
to the employers, and the employers pick people based on their proper qualifications, allowing
them to have higher paying jobs.
The above chart shows four things. Firstly, it shows that education causes income. This
idea is already supported by the examples above. Secondly, it shows that education causes
credentials. Thirdly, the chart explains that credentials cause income. And lastly, it shows
where education doesn’t lead to credentials, it cannot lead to higher income. st
SOC200H1—Measurement October 1 2012
When testing a hypothesis about a mechanism, you are testing to control for, or hold
constant, the intervening variable. For example, the first constant would be a high school
diploma, and the second constant would be the post-secondary degree. It should be noted that it
is not the years of education that maters, but rather the credentials. But how do we know if the
variable is spurious or if it is intervening? We find this out by using logic and theory. There is a
difference between spuriousness and the intervening variable. Spuriousness is a variable that
affects both the independent and dependent variable. The intervening variable is an independent
variable that causes an intervening variable, which then causes the dependent variable.
How does X affect Y under different circumstances?
Not only do things affect variables, but the relationships between variables can be affect
by things as well. If it’s an “it depends” situation, then it is called an interaction effect. An
example of this is that quality math skills result in good grades. Therefore, math skills will get
you higher grades in only math related courses, not in English. This constitutes a “it depends”
situation. Another way to look at this is that even though math skills equal good grades, this
only applies if you have a math major.
A second example of this logical is the idea that unemployment equals good health. This,
however, only depends on an individuals source of health insurance. Different health rates can
have different effects.
A third example is that higher levels of education equal lower levels of depression. This
is only true depending on gender .
A fourth example is the idea that Tylenol increases peoples quality of health and life.
This is dependent on many other variables. If the person is a chronic smoker or drinker, their life
expectancy will still be diminished. It also depends on what diet the individual has.
To determine how X affects Y and under what conditions, you must test for interactions.
This means to control for, or hold constant, the variable that you think might affect the
relationship. To do this, you must figure out the slope of the line on the graph. If it’s the same
slope between variables, then there is no interaction. It the slope is different, then there is an
interaction. For example, the income and occupation status between black and brown shoes. Or
the income status between men and women. Another example could also be the idea of health
and employment between public health insurance and employer heath insurance. Again, this is
dependent on the insurance you have. From this we can conclude that mechanisms can
sometimes suggest interacting effects.
No theory or hypothesis holds, even probabilistically, will hold in every place at every